Update app.py

app.py

@@ -37,12 +37,11 @@ def extract_text_from_images(image_paths):
 
 # Step 3: Rebuild the PDF with Extracted Text
 def rebuild_pdf(original_pdf_path, text_data, output_pdf_path):
-    """Overlay extracted text onto the original PDF."""
+    """Overlay extracted text onto the original PDF with enhanced alignment."""
     with fitz.open(original_pdf_path) as pdf:
         output_pdf = fitz.open()
 
         for page_num, page in enumerate(pdf):
-            # Render page as an image for reference
             pix = page.get_pixmap(dpi=300)
             img = Image.frombytes("RGB", [pix.width, pix.height], pix.samples)
 
@@ -54,15 +53,25 @@ def rebuild_pdf(original_pdf_path, text_data, output_pdf_path):
                 if conf > 0.6:  # Confidence threshold
                     (x_min, y_min), (x_max, y_max) = bbox[0], bbox[2]
                     rect = fitz.Rect(x_min, y_min, x_max, y_max)
+
+                    # Adjust text placement for alignment
+                    adjusted_x = x_min + 2  # Fine-tune as needed
+                    adjusted_y = y_min + 2  # Fine-tune as needed
+                    adjusted_rect = fitz.Rect(adjusted_x, adjusted_y, x_max, y_max)
+
                     new_page.insert_textbox(
-                        rect, text, fontsize=12, fontname="helv", color=(0, 0, 0)
+                        adjusted_rect,  # Use adjusted bounding box
+                        text,
+                        fontsize=10,  # Adjust font size as needed
+                        fontname="helv",  # Use Helvetica for compatibility
+                        color=(0, 0, 0),  # Black color for text
                     )
 
             # Add original diagrams and graphics
             new_page.show_pdf_page(page.rect, pdf, page_num)
 
-        # Save rebuilt PDF
-        output_pdf.save(output_pdf_path)
+        # Save rebuilt PDF with embedded fonts
+        output_pdf.save(output_pdf_path, garbage=4)
 
 
 # Full Workflow